We've had problems with a number of devices that communicate over power lines -- both the TED 5000 electricity monitor and most recently a pair of Netgear powerline networking adapters.

I've tried a first pass of trying to find a particular appliance that might be generating noise, but even with all the breakers off but one we still have the same problems.

I'm looking for suggestions as to what I can do to diagnose this problem...is there any way for me to diagnose the magnitude of "noise" on the circuits? Are there other things I should try?

  • Here's a related question: How do I troubleshoot noisy power lines for power-line networking?, though it looks like you've already done some of the things suggested there.
    – Niall C.
    Jan 7, 2011 at 15:08
  • 1
    I have tried most of that...on the other hand, we do have a bunch of CF lights, although that didn't appear to affect the problem one way or the other when I was trying to get the TED5000 working. I may have to repeat the whole turning-everything-off test this weekend with the powerline devices. I guess my real question is, other than plugging in an oscilloscope (which I don't own), are there any tools that would provide some sort of metrics about the noise level on the circuits?
    – larsks
    Jan 7, 2011 at 16:07

3 Answers 3


You don't specify what kind of problems you're experiencing. No signal at all? Low signal? Dropped packets?

The latest NEC codes require arc fault interrupters, which have been shown to interfere with powerline signals.

Further, as others have explained, the powerline signal may not bridge correctly to the other leg of the powerline. There are bridges that you can connect to both 120V legs of the 220V power input that will allow the signals to more easily pass from one leg to the other.

However, if you can't discover the issue, then you will likely need to take an oscilloscope to the main line to see if you are getting interference from your neighbors.

Lastly, high RF can impair these devices. If you are near transmission towers or other RF generating businesses then you may simply be out of luck.

  • "Low signal" or "no signal" probably describes the problem (the devices I'm using don't actually provide a signal strength indicator. The netgear devices have a "connection speed" indicator that I imagine is somewhat equivalent).
    – larsks
    Jan 12, 2011 at 15:44

120714-1952 EDT

TED power line communication is a moderately steady data stream. X-10 and similar devices are very intermittent.

TED 1000 outputs a stream (packet) of data about 0.1 second long, and this occurs every 1 second. Thus, a single TED MTU is consuming 10% of the channel capacity. One packet extends over about 6 cycles of the 60 Hz supply. There is only one way communication in the TED 1000 system for normal data collection.

TED 5000 outputs a stream of data about 0.2 seconds long, and this occurs every 2 to 5 seconds. Again for a single MTU it may be about 10% of the channel capacity. In addition every 1 minute a longer packet is sent that requires an acknowledgement from the receiving end.

In the TED system a lot of data is sent in 1 second. In the X-10 system one on or off command can be sent in about 1 second. For most applications the TED data only needs to go to one location to do an adequate job. In the X-10 application the command must go to many dispersed locations in a broadcast fashion. TED really does not need to operate in a broadcast type mode. However, TED is sold with that capability as a feature.

If I was designing a TED system I would use a dedicated wired connection. Too many customers are tolerant of computers and other electronic devices with bugs and other types of problems. PLC (power line communication) appears, from the comments on the Internet, to cause a lot of problems for TED customers.

A TED system operates at a 125 kHz carrier frequency. This is about the 7500th harmonic of 60 Hz. Unless you employed a very good square-wave shaping circuit on the basic 60 Hz sine wave there would be little energy at 125 kHz. There may be devices connected to the AC line and synchronized to 60 Hz that generate substantial power in the 125 kHz range, but I don't really want to call these harmonics. Rather I would consider them as uncorrelated noise.

What is a good reliable method to work with the TED system, and also prevent it from interfering with other systems? Select an existing 120 V branch circuit, or create a new circuit. This is to be an isolated circuit with only one RDU xor Gateway and the associated MTUs as the connections to the isolated branch circuit. Where this isolated circuit connects into the main panel place an X-10 in-line filter between the breaker and the isolated circuit. The in-line filter places a high impedance at around 125 kHz between the isolated circuit and all other circuits.

With this scheme I can get zero packet errors over very long time periods. No problem over a 250 ft length of #14 Romex cable.

You can find additional information at my web site http://beta-a2.com/ . Pick Energy for a summary of my book ( http://beta-a2.com/energy.html ), and pick TED MTU Data Format for a discussion on the 1000 and 5000 packet structures.



Many of these powerline networking devices cannot properly communicate across phases of the circuit breaker panel.

If all of your devices are on the same side (either left or right) of the panel, they'll be on the same phase. If one device is on the other side phase, it will likely have problems communicating with the others (if it can at all).

EDIT: See @Tester101's comment for a correction on what breakers are on which phases.

  • The two powerline devices are on the same phase. The solution to the TED5000 was using the outlet that was only a few inches from the panel (communicating with the device attached to both phases in the panel), so really the minimum possible length of travel.
    – larsks
    Jan 7, 2011 at 16:04
  • @larsks, I figured from your use of TED 5000 you probably knew to avoid the phase problem. Hopefully other readers who get here because of the Netgear networking adapter mention can find the answer useful though. ;-)
    – BQ.
    Jan 7, 2011 at 16:11
  • 2
    In the US phase is not denoted by the 'side' of the panel the breaker is on. It is by the row the breakers are in. For example 1 & 2 = Phase A, 3 & 4 = Phase B, 5 & 6 = Phase A, 7 & 8 = Phase B... and you'll notice that all the odds are on one side and evens on the other. See this wiki entry en.wikipedia.org/wiki/Distribution_board
    – Tester101
    Jan 7, 2011 at 17:41
  • @Tester101, +1 interesting. Didn't know that. I'd just taken that info from message boards when I was deciding what power monitoring solution to get. I'd seen the side/phase referenced several places and assumed it to be correct. I'll edit my answer to mention this.
    – BQ.
    Jan 7, 2011 at 19:15
  • 2
    Hey guys, I don't have a good suggestion to the noise problem without an oscilloscope sensitive down to 60HZ,(actually there are some fairly cheap handheld ones out there) and the process of elimination. i do want to do some vocabulary preaching, if I may. All 120/240VAC residential service in the USA is single phase. the two sides (120VAC each) are referred to as legs. A phase is a vector equivalent measured in degrees and is created my a delta or Wye transformer, 208VAC/480VAC. etc. Phase electricity is only found on industrial applications. Jan 11, 2011 at 21:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.